QVAC-20631 chore: DO NOT MERGE, overlay-validate the 7 fabric consumers against qvac-fabric 10297.1.0

[iancris]

DO NOT MERGE

Validation branch for the qvac-fabric 10297.1.0 rollout, Phase A. It carries a temporary vcpkg overlay that pins qvac-fabric to commit b098a1f0508356bbfd97c8ba761ebcdcdce682e2 — the head of fabric PR tetherto/qvac-fabric-llm.cpp#214, "QVAC-24112 fit: budget against real memory availability".

Merging this would point every consumer at a temporary overlay. The overlay is validation scaffolding only and must be removed in Phase B, after the registry port is published and the tag v10297.1.0 exists.

What problem does this PR solve?

The 7 fabric consumers need to be validated against the fabric commit that will become 10297.1.0 before the consumer dependency floors are bumped.

The overlay on this branch previously pinned the temp-10297 branch head (29db2fb9f). Fabric PR #214 is 5 commits ahead of, and 0 behind, temp-10297 and is not yet merged, so that pin never exercised it. #214 changes API surface the consumers compile against:

• ggml/include/ggml-backend.h — public header
• common/fit.cpp, common/fit.h — the budgeting API @qvac/model-fit is named for
• ggml/src/ggml-cpu/ggml-cpu.cpp, ggml/src/ggml-metal/ggml-metal-device.m, ggml/src/ggml-metal/ggml-metal.cpp

How does it solve it?

• Adds the shared overlay port at vcpkg-overlays/ports/qvac-fabric/, pinned by commit SHA (not a tag — v10297.1.0 does not exist yet).
• Adds "overlay-ports": ["../../vcpkg-overlays/ports"] to all 7 fabric consumers:
  • embed-llamacpp
  • fabric
  • llm-llamacpp
  • model-fit
  • ocr-ggml
  • translation-nmtcpp
  • vla-ggml
• Leaves consumer version>= pins untouched at 10297.0.0. The overlay bypasses version resolution entirely, so only the overlay port's own version matters here; bumping the pins is Phase B.
• Leaves default-registry.baseline untouched in all 7.

@qvac/classification-ggml is deliberately not included — it dropped the qvac-fabric vcpkg dependency and consumes the published npm package instead.

Tested

• Diff against the merge base is exactly the 9 expected files, 302 insertions, 0 deletions.
• SHA512 computed with shasum -a 512 over the /archive/ tarball vcpkg actually fetches (37,446,085 bytes, gzip -t clean). Worth confirming with vcpkg hash before final handoff, as vcpkg was not available in the authoring environment.
• All 7 on-pr-<consumer>.yml workflow files confirmed present on this ref.

Notes for reviewers

• Expect cpp-lint failures that are not this change. Re-pinning the overlay changes the qvac-fabric port's vcpkg ABI hash, so qvac-fabric is a guaranteed cache miss and gets compiled from source on every consumer. Compiling any port from source on the qvac-ubuntu2204-x64 pool currently trips a libc++/clang version mismatch in the runner image — tracked as QVAC-24201. The signature is "Libc++ only supports Clang 20 and later" followed by error: "remove_reference not implemented!". That is environmental, not fabric.
• Fabric PR release(qvac-lib-registry-client): v0.2.0 #214 is not merged, and its own CI has a failing windows job. If its head moves, this overlay must be re-pinned and all 7 consumers re-run.
• Earlier green results on this branch against the 29db2fb9f pin do not carry over — they validated a fabric that did not contain the fit change.

[github-actions]

Review Status

Current Status: ❌ PENDING
Approvals so far: none

Pending reviews: Needs 1 Management or Team Lead, and 1 more from Management, Team Lead, or Member.

[github-actions]

License compliance — clean

No new dependency license findings in this PR.

Warn-only (shadow) mode — this check does not block merges yet.

Updated automatically by the canonical license compliance workflow.

NOTICE presence (advisory)

Missing NOTICE (advisory, does not block):

• ./.github/actions/release-merge-guard
• ./docs/website
• ./packages/ggml-coload-smoke
• ./packages/fabric/test/integration
• ./packages/inference-addon-cpp/mobile
• ./packages/sdk/e2e
• ./packages/llm-llamacpp/benchmarks/performance
• ./packages/llm-llamacpp/benchmarks/server
• ./packages/vla-ggml/sim/server
• ./packages/embed-llamacpp/benchmarks/performance
• ./packages/embed-llamacpp/benchmarks/server
• ./packages/asr-ggml/benchmarks/server

[github-actions]

🧪 C++ Test Coverage Report

Coverage:

📊 Detailed Coverage

Filename                         Regions    Missed Regions     Cover   Functions  Missed Functions  Executed       Lines      Missed Lines     Cover    Branches   Missed Branches     Cover
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
NmtLazyInitializeBackend.cpp          99                20    79.80%          11                 1    90.91%         157                36    77.07%          66                21    68.18%
NmtLazyInitializeBackend.hpp           2                 0   100.00%           1                 0   100.00%           1                 0   100.00%           0                 0         -
TranslationModel.cpp                 296               168    43.24%          28                 8    71.43%         506               213    57.91%         181               122    32.60%
TranslationModel.hpp                   1                 0   100.00%           1                 0   100.00%           1                 0   100.00%           0                 0         -
nmt.cpp                               72                22    69.44%           9                 1    88.89%         137                28    79.56%          44                16    63.64%
nmt.hpp                               51                 4    92.16%          11                 2    81.82%          53                 4    92.45%          28                 0   100.00%
nmt_beam_search.cpp                  116                25    78.45%          10                 3    70.00%         254                32    87.40%          76                19    75.00%
nmt_graph_decoder.cpp                164                78    52.44%          15                 7    53.33%         540               161    70.19%         112                69    38.39%
nmt_graph_encoder.cpp                 54                13    75.93%           3                 0   100.00%         268                33    87.69%          37                16    56.76%
nmt_loader.cpp                       270                67    75.19%          14                 0   100.00%         774                97    87.47%         161                67    58.39%
nmt_state_backend.cpp                253                94    62.85%          21                 0   100.00%         489               128    73.82%         165                87    47.27%
nmt_tokenization.cpp                  88                21    76.14%           8                 0   100.00%         135                36    73.33%          61                26    57.38%
nmt_utils.cpp                        120                89    25.83%           8                 3    62.50%         180               134    25.56%          78                63    19.23%
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
TOTAL                               1586               601    62.11%         140                25    82.14%        3495               902    74.19%        1009               506    49.85%